There is an ever increasing need for clean and safe air to breathe and water to drink, particularly in heavily populated countries or regions throughout the world. A major, high-volume, application for compact solid-state deep UV light sources is for chemical-free sterilisation of air or water. Deep UV light—that is light in the UVC range which has wavelength shorter than 280 nm—efficiently causes permanent physical damage to DNA which prevents bacteria, viruses and fungi from replicating. This means that deep UV treatment can be used to disinfect air or water at point-of-use for safe breathing or drinking. Deep UV light is particularly effective at destroying e-coli bacteria. Deep UV light can also be used to disinfect surfaces.
Deep UV light can also be used to reduce the toxicity of chemical pollutants such as dissolved organic chemicals which are present in water and thereby make the water safe to drink. In this case the deep UV light initiates photocatalytic oxidation reactions which break down the dissolved organic chemicals into less-hazardous or non-hazardous byproducts. The initiation of photocatalytic oxidation reactions is most efficient for deep UV light with wavelengths shorter than 230 nm.
Compact solid-state deep UV light sources also have application in bio- and chemical-sensing because biological and chemical compounds strongly absorb deep UV light. Proteins and other organic chemicals can be identified from their fluorescence spectra. A fluorescence measurement requires illumination with light at a short wavelength at which the compounds are strongly absorbing and detection of the resulting fluorescence at longer wavelengths. Wavelengths near 280 nm are suitable but shorter 220 nm wavelengths are much preferred owing to the stronger absorbance at this wavelength.
Point-of-use products for the UV treatment of air and water are already available and these products use mercury lamps as the UV light source. However, mercury lamps contain toxic material, tend to have short operating lifetimes and long warm-up times and require high driving voltages. Furthermore, the UV light emitted from mercury lamps is emitted in a broad range of directions and from a relatively large area which means it cannot be efficiently focused into a small area or a collimated beam.
An alternative UV light source currently under development is the UV LED. The current draw-backs to using UV LEDs include that they have short operating lifetimes and that they cannot be efficiently focused to provide a collimated beam or tightly focused light spot. In addition, the performance of UV LEDs with emission wavelengths shorter than 260 nm is very poor. Therefore, these devices are poorly suited for the applications described above which benefit from a light source with wavelength shorter than 260 nm.
Deep UV lasers potentially provide a monochromatic, coherent beam which can be efficiently collected into a collimated beam or focused into a small area and can be modulated rapidly (as required for fluorescence measurements, for example). However, existing lasers with emission wavelengths shorter than 280 nm are very expensive components such as gas lasers designed for industrial use. No laser diodes have been made with emission wavelength shorter than 280 nm.
A deep UV laser can be realised by frequency doubling a visible laser beam inside a suitable non-linear optical material. (e.g. Beta-Barium Borate which is commonly known as BBO), as first reported in IEEE Journal of Quantum Electronics QE-22, No 7 (1986). The visible light is focused into the non-linear optical material and the light is frequency-doubled (FD) by the process of second harmonic generation (SHG). The SHG process converts the visible input light into light with wavelength half the wavelength of the input light. The frequency-doubled light has properties similar to the properties of light emitted by a laser and it is common in the prior art for the light to be described as “laser light”. It is also common in the prior art for a device which emits frequency-doubled light to be described as “laser”. As used herein, the term “laser light” includes light emitted by a laser device as well as frequency-doubled light derived from light emitted by a laser device. Further, as used herein a “laser light source” includes a light source that exhibits light amplification by stimulated emission of radiation, as well as a device that that implements frequency doubling of light emitted by a light source that exhibits light amplification by stimulated emission of radiation. Frequency-doubled UV lasers made in this way using BBO can be made to emit wavelengths as short as 205 nm.
Nishimura et al in JJAP 42, 5079 (2003) were the first to report on making a UV laser using BBO and using a blue-violet semiconductor laser diode to generate the “pump” visible laser beam. A potential advantage of this approach is that blue-violet semiconductor laser diodes are compact and low-cost components. However, the SHG process in BBO occurs with low efficiency for the relatively low powers of light emitted from blue-violet semiconductor laser diodes. Consequently, although a complex optical system was used to recirculate the blue light through the BBO component to improve efficiency, the UV output power achieved in this prior art was still low. Therefore, this method is not suitable to fabricate a low-cost high-power UV light source for the applications described above. Neither of the two most recently mentioned pieces of prior art discuss the use of a frequency doubling waveguide.
One method which has been used to increase frequency-doubling efficiency has been to use a frequency-doubling waveguide. A frequency-doubling waveguide is designed to confine the pump light and frequency-doubled light to a small cross-sectional area as they pass through the non-linear optical material. The light is confined and guided along the non-linear optical material by internal reflection at the interfaces between the non-linear optical material and the surrounding material (or gas), which have a different refractive index. The light may be confined in one dimension which is perpendicular to the propagation direction of the light—this is commonly referred to as a “planar” waveguide. Alternatively the light may be confined in the two dimensions which are perpendicular to the propagation direction of the light in either “channel” or “ridge” waveguides. By confining the light to a small area the efficiency of the SHG process can be significantly increased. The earliest report of using a frequency doubling waveguide is disclosed in U.S. Pat. No. 3,584,230 (Tien, Jun. 8, 1971) in the form of a thin non-linear optical film deposited on a substrate. This prior art does not use a visible laser diode nor BBO as the FD waveguide material so does not provide a method to make a UV laser. In particular, there is no method known in the prior art to deposit high-quality single crystal BBO thin films onto substrates, other than homoepitaxial deposition onto BBO substrates which provide no refractive index contrast with the deposited BBO layer as required for a waveguide. Therefore, this method is not suitable to fabricate a high-quality waveguide for a deep UV laser.
A waveguide can be formed inside a bulk non-linear optical crystal by generating a refractive index contrast within the crystal using conventional methods of diffusion, proton exchange or implantation. U.S. Pat. No. 4,427,260 (Puech et al., Jan. 24, 1984) describes an invention of a non-linear optical device where a laser diode pumps a FD waveguide formed using Ni diffusion. This prior art does not discuss the use of BBO, nor any other nonlinear material suitable for frequency-doubling to deep UV wavelengths, and it is not clear that diffusion can be used to form high-quality waveguides suitable for use in a deep UV laser. Furthermore, this prior art does not provide cladding layers with composition significantly different from that of the non-linear crystal, as is required to make a waveguide where the light is strongly confined. APL 41, 7, p 607 (1982), U.S. Pat. No. 4,951,293 (Yamamoto et al., Aug. 21, 1990), and APL 85, 9 1457 (2004) report the formation of FD waveguides using Ti diffusion, proton exchange or implantation. The latter reports on using BBO but none of them provide cladding layers with composition significantly different from that of the non-linear crystal. Furthermore, these methods for fabrication a waveguides tend to result in high absorption losses for ultraviolet light with very short wavelength (e.g. wavelength less than 280 nm).
U.S. Pat. No. 5,175,784 (Enomoto et al., Dec. 29, 1992) describes a FD waveguide structure made by depositing a non-linear optical thin film onto a substrate and then etching into a ridge structure. BBO is given as several examples. However, there is no method known in the prior art to deposit high-quality single crystal BBO thin films onto substrates, other than homoepitaxial deposition onto BBO substrates which provide no refractive index contrast with the deposited BBO layer as required for a waveguide. Therefore, this method is not suitable to fabricate a high-quality waveguide for a deep UV laser.
APL 89 041103 (2006) reports on the formation of a frequency doubling ridge waveguide in a BBO crystal using implantation of helium ions and dry etching. The generation of a UV laser by FD a visible laser beam using the waveguide is also reported. The implanted helium ions form a thin layer a few micrometers below the top surface of the crystal which has a slightly lower refractive index than that of the crystal between the layer and the surface. The light is confined in the crystal between the top surface and the implanted layer. There are significant disadvantages to using implantation to form the waveguide. In particular, the refractive index contrast between the implanted layer and the BBO crystal is relatively weak, which means light can “leak” out of the waveguide, the BBO crystal between the implanted layer and the surface is damaged during the implantation process and this reduces the UV output power, and absorption losses for ultraviolet light with very short wavelength (e.g. wavelength less than 280 nm) tend to be high. The visible laser used with the waveguide in this prior art was a bulky and expensive industrial laser which is a further disadvantage of this method.
The use of a lapping and polishing process to fabricate a thin film FD waveguide from a bulk crystal of non-linear optical material is described in two pieces of prior art. Deglinnocenti in PhD dissertation ETH No 17145 (2007) mentions that a thin film waveguide can be made by polishing down BBO crystals glued or optically mounted on fused silica substrate. This proposal neglects an important consideration in waveguide design—that of minimising optical absorption losses of the frequency-doubled light. An efficient waveguide requires that the materials surrounding the non-linear optical material core of the waveguide have low absorption of the light in the waveguide. Most materials are strongly absorbing of deep UV light (i.e. wavelengths shorter than 280 nm). In particular this is the case of the vast majority of glues and mounting materials which would be used to attach the BBO crystal to the fused silica substrate. Consequently, a BBO thin film fabricated according to the proposal in this prior art will have high absorption losses of the UV laser wavelength and therefore provide low efficiency. Furthermore, the absorption of the deep UV light will likely contribute to degradation of the material and result in a short lifetime of the waveguide component. U.S. Pat. No. 6,631,231 (Mizuuchi et al., Oct. 7, 2003) discloses an optical waveguide element made by gluing an FD crystal to a substrate and where the glue acts as a cladding region. No mention of BBO is made in this prior art nor is the generation of deep UV laser light.
U.S. Pat. No. 5,123,731 (Yoshinaga et al., Jun. 23, 1992) discloses a laser source that emits both a frequency doubled UV laser beam and another laser beam generated by a laser diode. The use of a frequency doubling waveguide is also disclosed. This prior art does not mention using a visible laser diode nor does it give any details on the construction method of the waveguide.